Runway vs. Taxiway: Challenges in Automated Line Identification and Notation Approaches
Parth Ganeriwala, Amy Alvarez, Abdullah AlQahtani, Siddhartha Bhattacharyya, Mohammed Abdul Hafeez Khan, Natasha Neogi · 2025
Read Runway vs. Taxiway: Challenges in Automated Line Identification and Notation Approaches by Parth Ganeriwala, Amy Alvarez, Abdullah AlQahtani, Siddhartha Bhattacharyya, Mohammed Abdul Hafeez Khan, Natasha Neogi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-18494.txt.
Introduction
Parth Ganeriwala and co-authors try to teach software to tell runways from taxiways. Their 2025 paper deals with auto labeling of airfield lines, a safety base for landing aids and ground control. A stock algorithm aced taxiway lines but stumbled on runways, whose paint, setting, and wear differ. Shadows, tire marks, and shifting light made the gap worse.
Tweaks to color limits and search regions bought only small gains, with the horizon still misread as a line. Their fix adds a neural classifier, AssistNet, as a check step in the chain. The authors frame the pairing as a sturdier path to trusted airfield labels.
Catalog record
| CREATOR | Parth Ganeriwala, Amy Alvarez, Abdullah AlQahtani, Siddhartha Bhattacharyya, Mohammed Abdul Hafeez Khan, Natasha Neogi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-18494
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Runway vs. Taxiway: Challenges in Automated Line Identification and Notation Approaches Parth Ganeriwala, Amy Alvarez, Abdullah AlQahtani, Siddhartha Bhattacharyya, Mohammed Abdul Hafeez Khan, Natasha Neogi Abstract The increasing complexity of autonomous systems has amplified the need for accurate and reliable labeling of runway and taxiway markings to ensure operational safety. Precise detection and labeling of these markings are critical for tasks such as navigation, landing assistance, and ground control automation. Existing labeling algorithms, like the Automated Line Identification and Notation Algorithm (ALINA), have demonstrated success in identifying taxiway markings but encounter significant challenges when applied to runway markings. This limitation arises due to notable differences in line characteristics, environmental context, and interference from elements such as shadows, tire marks, and varying surface conditions. To address these challenges, we modified ALINA by adjusting color thresholds and refining region of interest (ROI) selection to better suit runway-specific contexts. While these modifications yielded limited improvements, the algorithm still struggled with consistent runway identification, often mislabeling elements such as the horizon or non-relevant background features. This highlighted the need for a more robust solution capable of adapting to diverse visual interferences. In this paper, we propose integrating a classification step using a Convolutional Neural Network (CNN) named AssistNet. By incorporating this classification step, the detection pipeline becomes more resilient to environmental variations and misclassifications. This work not only identifies the challenges but also outlines solutions, paving the way for improved automated labeling techniques essential for autonomous aviation systems. I Introduction The development of fully autonomous vehicles has made significant advancements in computer vision and robotics. Camera-based lane and line detection systems play a crucial role in enabling vehicles to perceive the surroundings and navigate safely [ 8 , 18 ] . Within the aviation industry, accurate runway detection is paramount for safe landings and takeoffs, particularly under challenging conditions like poor weather and low visibility. Accidents occurring during the taxiing phase, often attributed to factors like crew distractions and communication breakdowns with Air Traffic Control (ATC), highlight the need for automated perception systems that can augment pilot awareness and decision making on the ground. While some research has explored light-based guidance systems and other sensor modalities (e.g., LIDAR), computer vision-based approaches for taxiway line detection remain limited. Early efforts focused on grayscale images and nose-wheel mounted cameras [ 4 ] , while more recent work has incorporated color images [ 13 ] and explored camera placement on the aircraft’s vertical stabilizer [ 3 ] . However, these approaches have been primarily tested in simulated environments or under ideal conditions, and their performance in real-world scenarios remains a challenge [ 16 ] . This research builds on the work of Ganeriwala et al. [ 17 ] and Khan et al. [ 21 ] by enhancing the automated labeling process, specifically through the ALINA annotation framework [ 21 ] . ALINA was developed to detect and label taxiway line markings in video frames by establishing a consistent trapezoidal region of interest (ROI) in the initial frame, which is maintained throughout the video sequence. After applying geometric adjustments and color space transformations, ALINA generates a binary pixel map. The framework utilizes the CIRCular threshoLd pixEl Discovery And Traversal (CIRCLEDAT) algorithm to identify pixels representing taxiway markings, producing annotated frames and coordinate data files. ALINA was validated on 60,249 frames from the AssistTaxi dataset [ 17 ] , focusing on three videos with different camera perspectives, establishing a rigorous foundation for ALINA’s effectiveness and its potential scalability to larger datasets. However, while ALINA has demonstrated success with taxiway line detection, it has not yet been adapted for runway markings. Runway and taxiway line characteristics differ significantly in terms of environmental context, marking types, and structural layouts. This paper addresses the challenges in extending automated line detection algorithms to runway markings, highlighting the need for a classification step to distinguish between runway and taxiway surfaces before applying labeling algorithms. Initial modifications to ALINA, including adjustments to color thresholds and manual ROI selection, have shown limited improvements, with the algorithm often mislabeling background elements such as the horizon. Future adaptations, such as dynamic ROI adjustment and integration of contextual cues like horizon and runway geometry, are proposed to improve runway marking detection accuracy and robustness. The rest of this paper is structured as follows: Section II reviews related work, covering advancements in road and taxiway line detection, runway detection, and methods for horizon detection and contextual understanding. We also outline the challenges in runway marking detection, highlighting key limitations in current approaches and the need for improved classification methods. Section III discusses the initial experiments performed to adapt the ALINA framework for runway marking detection while section IV introduces our proposed classification approach, detailing the dataset, model architecture, and methodology used. Section V presents the results and discussion, with a case study on a subset of the AssistTaxi dataset. Finally, Section VII concludes the paper, summarizing the key findings and discussing future directions for adapting automated line identification in aviation contexts. II Related Works II-A Road and Taxiway Line Detection Existing research on road and taxiway line detection has primarily utilized geometric and model-based approaches, which leverage lane marking geometry through techniques like the Hough transform or parallelogram-based ROIs for line detection and tracking [ 2 , 20 ] . Feature-based methods, relying on characteristics like color, texture, and gradients, have been enhanced through machine learning techniques such as Support Vector Machines (SVM) and Principal Component Analysis (PCA) to improve detection accuracy [ 7 , 11 ] . Recent work has also explored deep learning approaches using convolutional neural networks (CNNs) and capsule networks, showing promise in high-resolution aerial imagery [ 19 ] . However, adapting these approaches for aviation applications presents unique challenges due to differences in line types and environmental variability. Meymandi-Nejad et al. [ 23 ] demonstrated that algorithms originally developed for automotive applications, such as particle filters, Hough transforms, and LaneNet [ 24 ] , are less effective in aeronautic environments, which include visual noise, variable lighting, and limited ROI defined by aircraft camera positioning. Delezenne et al. [ 10 ] extended this research by implementing a sliding window mechanism combined with airport map data, facilitating centerline identification and decision-making for autonomous UAV taxiing. This approach highlights the role of contextual data and specialized algorithms in achieving reliable taxiway navigation. Building on these insights, recent frameworks, including ALINA [ 21 ] , focus on the unique demands of aviation environments, leveraging contextual information and algorithmic adaptations to improve taxiway marking detection. The need for tailored solutions in aeronautics, including advanced vision-based navigation systems that can handle real-world variabilities, remains essential for future advancements in both manned and unmanned aviation ground operations. (a) Accurate detection of edge lines but missing the center lines (b) Partial detection achieved but with some missed segments due to propeller interference (c) Partial detection achieved but with some missed segments. (d) ALINA labels only the right line marking, with the left line missing Fig. 1 : Results of ALINA’s Line Labeling on Runway Instances with Adjusted Color Thresholding. II-B Runway Detection Runway detection research is less developed compared to road and taxiway line detection, with most approaches focusing on onboard camera systems